Hi, 03/07/2016 07:55 PM tarihinde Rahul Rege yazdı: > Hi, > > While this is not a question specifically for LXC/LXD, I thought I'd > ask to get the general opinion about the multi-tenancy and its support > in LXD (the nova-compute for same already gives some good ground). > > Its my understanding from the general reading that the Linux kernel is > not made for multitenancy, so if I am to implement a solution with say > LXD to run on baremetal servers, ideally I cannot put different > customers containers on the same host because of the potential > security issues (I fully don't understand what those are) > Why do you think lxc/lxd has of the potential security issues?
Please look at proxmox container! http://www.proxmox.com/en/proxmox-ve/features It use lxc for container. I am thinking lxc + apparmor + cgroups + lvm or raw image file(for container file system) are enough for security. we know setup is not easy therefore we use proxmox for lxc management . :-) If you have experience with " lxc + apparmor + cgroups + lvm or raw image file(for container file system)", you can try proxmox. > While different security techniques like Apparmor, selinux and some > capability restriction can achieve what we want ultimately, I wanted > to understand if people are generally doing it with LXD or would > potentially do it OR ultimately it'd follow the pattern of running > them on the VMs which provide the needed isolation. > > Companies like Joyent with their Triton containers do it I believe and > claim some really awesome performance with their SmartOS as the > special os. > > PS - This might be a well known consideration in container world, so > pardon me if this is something trivial and already discussed. > > > Thanks > Rahul Rege > > > > > _______________________________________________ > lxc-users mailing list > [email protected] > http://lists.linuxcontainers.org/listinfo/lxc-users
_______________________________________________ lxc-users mailing list [email protected] http://lists.linuxcontainers.org/listinfo/lxc-users
